[igt-dev] [PATCH i-g-t 3/3] tests/intel/freq: Test throttle basic API infra
Sundaresan, Sujaritha
sujaritha.sundaresan at intel.com
Wed Dec 6 04:44:36 UTC 2023
On 12/6/2023 3:08 AM, Rodrigo Vivi wrote:
> For now we don't have a reliable way to force PCODE throttling,
> but let's at least check for some sanity on our exposed API.
>
> Cc: Sujaritha Sundaresan <sujaritha.sundaresan at intel.com>
> Signed-off-by: Rodrigo Vivi <rodrigo.vivi at intel.com>
> ---
> tests/intel/xe_gt_freq.c | 48 ++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 48 insertions(+)
>
> diff --git a/tests/intel/xe_gt_freq.c b/tests/intel/xe_gt_freq.c
> index 719852530..650c54005 100644
> --- a/tests/intel/xe_gt_freq.c
> +++ b/tests/intel/xe_gt_freq.c
> @@ -68,6 +68,49 @@ static uint32_t get_freq(int fd, int gt_id, const char *freq_name)
> return freq;
> }
>
> +static uint32_t get_throttle(int fd, int gt_id, const char *throttle_file)
> +{
> + uint32_t val;
> + char throttle_attr[40];
> + int gt_fd;
> +
> + snprintf(throttle_attr, sizeof(throttle_attr),
> + "freq0/throttle/%s", throttle_file);
> + gt_fd = xe_sysfs_gt_open(fd, gt_id);
> + igt_assert(gt_fd >= 0);
> +
> + igt_sysfs_scanf(gt_fd, throttle_attr, "%u", &val);
> +
> + igt_debug("gt%d/freq0/throttle/%s: %u\n", gt_id, throttle_file, val);
> +
> + close(gt_fd);
> + return val;
> +}
> +
> +/**
> + * SUBTEST: throttle_basic_api
> + * Description: Test basic throttle API
> + */
> +
> +static void test_throttle_basic_api(int fd, int gt_id)
> +{
> + uint32_t status, reasons;
> +
> + status = get_throttle(fd, gt_id, "status");
> + reasons = get_throttle(fd, gt_id, "reason_pl1");
> + reasons |= get_throttle(fd, gt_id, "reason_pl2");
> + reasons |= get_throttle(fd, gt_id, "reason_pl4");
> + reasons |= get_throttle(fd, gt_id, "reason_prochot");
> + reasons |= get_throttle(fd, gt_id, "reason_ratl");
> + reasons |= get_throttle(fd, gt_id, "reason_thermal");
> + reasons |= get_throttle(fd, gt_id, "reason_vr_tdc");
> + reasons |= get_throttle(fd, gt_id, "reason_vr_thermalert");
> +
> + if (status)
> + igt_assert(reasons);
> + else
> + igt_assert(!reasons);
> +}
Hi Rodrigo,
Some weird alignment issues above ?
Otherwise LGTM.
Reviewed-by: Sujaritha Sundaresan <sujaritha.sundaresan at intel.com>
>
> /**
> * SUBTEST: freq_basic_api
> @@ -289,6 +332,11 @@ igt_main
> stash_max = get_freq(fd, 0, "max");
> }
>
> + igt_subtest("throttle_basic_api") {
> + xe_for_each_gt(fd, gt)
> + test_throttle_basic_api(fd, gt);
> + }
> +
> igt_subtest("freq_basic_api") {
> xe_for_each_gt(fd, gt)
> test_freq_basic_api(fd, gt);
More information about the igt-dev
mailing list